Re: Options for dealing with IDs

On Friday, January 10, 2003, 4:29:20 PM, Jeni wrote:

JT> Hi Chris,

>> On Thursday, January 9, 2003, 1:50:54 PM, Robin wrote:
>> RB> Another issue is that they are normally (in all cases I've seen
>> RB> them used in at least) sensitive to the default namespace.
>> Sensitive in that they can't use it?

JT> Sensitive in that they *do* use it. If, in XML Schema, you have:

JT> <schema xmlns="">
JT>   <element name="foo" type="string" />
JT> </schema>

JT> the type attribute holds an QName; the default namespace is used when
JT> interpreting this QName; since the value of the type attribute doesn't
JT> have a prefix, the processor uses the default namespace and recognises
JT> the value of the type attribute as being
JT> {}string.

Okay; so how do I declare an attribute that can take a Qname as its
attribute value, when I want that value to be an attribute name which
uncludes unprefixed attributes? Its seems from what you are saying that
I can't, which is a problem. A big problem, given that the majority of
attribute in practice are not namespaced.

JT> This differs from XPath where the default namespace is *not* used to
JT> resolve prefix-less QNames.

That was what I was thinking of when I wrote that it was not used. Oh
joy, two different ways of resolving the same thing.

JT> It sounds as if you're saying that the default namespace would not
JT> apply when interpreting the value of the xml:idAttr attribute.


JT> This means that the value of the xml:idAttr is not an xs:QName
JT> (per XML Schema's definition), but a NameTest (per XPath's
JT> definition).

Thanks for the warning about Schemas use of QNames. OK, pity, I would
rather have used W3C XML Schema definitions throughout.


Received on Friday, 10 January 2003 11:21:25 UTC